Professor Robert Shiller:
I wanted to talk today about insurance,
which is another risk management device that’s
traditionally separate from securities,
which we talked about last time, but the underlying
principles are the same. Before I begin,
I want to just give some more thoughts about the
diversification through securities and that will lead us
into insurance. Let me just review the
preceding lecture briefly for that purpose.
What we did–the core theoretical framework that we
had–was the mean variance theory, which led us to the
capital asset pricing model. But the basic thing was that we
had to–in order to use the framework–we had to start by
producing estimates of the expected returns on each asset,
we called those r, and the standard deviation of
the return on each asset and the covariance between the returns
of each pair of assets. Then, once we did that we could
plug that into the formula that I gave you last time and get the
standard deviation of the portfolio and the expected
return on the portfolio. From then on,
if you accept the analysis and the assumptions or the estimates
that underlie it, then we pretty much know how to
construct portfolios. The underlying estimates may
not accord with your belief or your intuitive sense of common
sense. The other thing that I
mentioned last time was that there seems to be a really big
difference between the expected return on the stock market and
the expected return on short-term debt.
We found an equity premium–or actually Jeremy Siegel’s book
gave an equity premium of 4% a year.
Some people find that hard to believe.
How can it be that one asset does 4% a year better than
another? Some people say,
well if that’s the case I want to invest in nothing more than
that one asset. Why should I take something
that is underperforming? Jeremy Siegel goes on further
to say that since the mid-nineteenth century we’ve
never had a thirty-year period when stocks under performed
bonds, so stocks are really–if anyone
who has an investment horizon of thirty years–you’d think,
why should I ever holds bonds. The numbers that Jeremy Siegel
produces seem implausibly high for the stock market.
What we call this is the–I want to emphasize it,
I’ll write this again–the equity premium puzzle.
That term was actually coined by economists,
Prescott and Mehra; it’s now in general use.
That is, it just seems that stocks so much outperform other
investments. For Jeremy Siegel,
in the latest edition of his book, the equity premium is 4% a
year since 1802. That’s almost–no that’s more
than 100–that’s 206 years. Why would that be and can you
believe that? One question that comes up is
that maybe–this is for the U.S. data–and some people say,
well, maybe, why are we looking at the U.S.?
Because the U.S. is an arguably very successful
country, so we have, potentially,
a bias in–it’s called a selection bias.
If you pick as the country you study one of the most successful
countries in the world, that doesn’t inform you very
well about what it is for a random country or for the U.S.
going forward, there’s something wrong.
The U.S. has been successful in
financial markets and it’s being imitated by lots of countries.
Financial markets similar to ours are being set up in many
places. You wonder, you know,
maybe they’re over imitating; maybe we were just lucky or
maybe it was because the U.S. was the first,
in some ways, to develop some of these
financial institutions–or one of the first.
But now, when more and more countries start doing it,
maybe it won’t work so well. One way of investigating this
is–to get around the selection bias–is to try to look at all
countries. Let’s not just look at the
United States; let’s look at every country of
the world and let’s see if they have an equity premium.
There’s a problem with that and the problem is that countries
that are less successful don’t keep data–that’s a problem.
Or they–sometimes they just shut down their stock markets at
some point. This is since 1802–now how
many countries do you think have uninterrupted stock market data
since 1802? What do you think?
Name another country that probably has it.
What’s that? England, UK?
If you go onto the continent, though, they tended to be
interrupted by World War I and World War II.
What about Japan, do they have–do you think they
have uninterrupted? They had a little bit of a
problem around World War II and you can try to bridge the gap,
but–anyway, there are people who have tried
to sort this out. There’s one,
it’s a book by Dimson, Marsh, &
Staunton that–called Triumph of the
Optimists–that Jeremy Siegel quotes.
He has a table in the new, fourth edition of his book.
Dimson, Marsh and Staunton look at the following countries:
Belgium, Italy, Germany,
France, Spain, Japan, Switzerland,
Ireland, Denmark, Netherlands,
UK, Canada, U.S., South Africa,
Australia, and Sweden. Every one of them has a
positive equity premium; although the U.S.
is on the high side of them all, it’s not the best.
The country that has the highest equity premium–and
that’s for the whole twentieth century,
they couldn’t go back to 1802–the most successful
country is Sweden and after that Australia.
U.S. is not the most successful
stock market although it’s high on the list.
Jeremy Siegel concludes that there’s–that the equity–he
said that these–that this book by Dimson,
Marsh and Staunton puts to bed any concerns about selection
bias and he claims that so many countries have shown an equity
premium that we can be confident.
His book is really very strong on the conclusions.
The title of the book, Stocks for the Long
Run–stocks always outperform other investments for
the long run and he says it’s not due to selection bias.
You know, I kind of wonder, the list of countries that I
just read to you, that Dimson,
Marsh and Staunton studied, excludes some important
countries, doesn’t it? Who does it exclude?
Well, it doesn’t have India, Russia, and China in it,
for example. At least Russia and China–do
you know anything about their history?
They have any stock market disruptions in the last one
hundred years? That’s kind of obvious.
They had a communist revolution in both places,
right? Russia and China are not
mentioned by–or not studied by Dimson, Marsh and Staunton.
Why not? Well, they can’t get data,
there wasn’t a stock market. Well there actually was a stock
market in Russia before 1918 and in China before 1949,
so what happened to investors? If you were a Chinese investor
in Chinese stocks in 1949, what happened?
We know what happened. It went–that’s that famous
minus 100% return, right, which dominates
everything. I think–what would Siegel say?
He’s really saying that this equity premium is enduring and
we should believe it. I don’t know,
I think that–I think what Jeremy would say is,
well you’re looking–if you look at Russia and China,
you’re looking at political factors and I’m only looking at
politically stable countries, so this whole thing is
irrelevant. Really, we’re not going to have
a communist revolution in any of these advanced countries now.
So Jeremy would say, forget that,
it looks pretty sound that we have an equity premium so we can
trust that. Well, he’s a good friend of
mine, but I think he may be overstating it a little bit;
we have some disagreements. The thing that comes to my mind
is that–I want to say before concluding this review of the
last lecture–that is that the stock market is inherently
political in any country. Politics have tremendous
effects on the values in the stock market and that’s because
of–even if the government doesn’t nationalize the stock
market or confiscate assets, they tax them.
Do you know we have, in the U.S.,
a corporate profits tax? Well, it’s not just in the
U.S., essentially every–I don’t know if there’s any exception.
There may not be an exception, but essentially every country
has a corporate profits tax and then we also have a personal
income tax. The corporate profits tax goes
after the profits that corporations make.
The personal–it’s taken from corporations before they pay out
their dividends. The personal income tax is
levied on individuals and these individuals have to pay it.
The personal income tax is not simple;
it’s not just a flat rate on your income, it depends on the
type of income. Dividend income or capital
gains income in the stock market is taxed differently.
The interesting thing is that through time,
as political winds change, these taxes have changed and
they’ve gone up to some very high levels in the past in the
United States and other countries.
I’m going to give some U.S. tax rates.
The personal tax on dividends–of course it depends
also on your tax bracket and your income;
I’m going to talk about the highest tax bracket.
In the U.S., it went over 90% in World War
II and the succeeding years. The government was taking 90%
of your dividend income. What is it today?
Does anyone know? What’s the tax rate of
dividends today? It might be zero for some
people, but it’s actually–it is–the standard rate for people
who have not negligible income is 15%.
It’s gone down from over 90% to 15%.
Why did it do that? Well, it’s some kind of
political change and the corporate–incidentally,
at the beginning of the twentieth century you were
right. Who said zero?
We didn’t even have income tax until 1913 when the Supreme
Court allowed it, so it was zero,
then it went up to 90%–or actually it was 94% at the
peak–and it came down to 15%. That’s a pretty big hit on the
stock market. So, it wasn’t just China that
took the stock market. When we were taking 90% of
dividends that was 90% of the stock market being taken by the
government; but that’s not all because we
were also taxing the corporations.
In the early post-war period, the corporate–now I’m going to
talk–there’s a distinction between the rate that they
charge and the actual amount that they take.
Most advanced countries of the world today have a corporate
profits tax rate for large corporations of about a third.
So they–a typical advanced country takes a third of the
profits, the government takes a third of the profits.
That’s not the actual amount that they pay because the tax
law is so complicated and there are so many loopholes.
What I looked at–and I have this on the website,
I have a chart showing corporate profits taxes paid,
as a fraction of corporate profits for the United States
since 1929. That has moved around a lot,
but it got almost up to 60% in the post-World War II period and
now it’s down to less than a third.
Why is it down? It’s because they’re changing
enforcement of the taxes and changing amounts of loopholes,
so most countries have a tax rate of about a third,
but corporations are paying less than a third of their
profits to taxes. If we want to look going
forward at the equity premium, we have to know how
much–what’s the politics? And what’s the government going
to do in the future? They’ve moved these tax rates
around a lot, so I think that it’s very hard
to be sure that we know going forward what are σ,
the expected return, the standard deviation of
returns and the covariance of returns–are really.
We have a nice theoretical framework, but the application
of the framework to the real data is hard and it ends up with
politics underneath it all; that’s just the real world.
I want to say one more thing about the diversification and
the mutual fund. Ideally, mutual funds are
calculating r and σ and σ_12 and
plugging in and finding what the optimal portfolio that you
should hold is, then offering that to you in
their mutual fund. Firms that do that are,
in practice, however, the minority and most
mutual funds have some gimmick or some special–they claim to
be beating the market not forming the optimal portfolio.
I also have up on the website some questions that I asked
investors about what they think about picking stocks.
Picking stocks means trying to find a stock that’s going to do
really well. What I found–the question I
asked is, do you agree–which of the following is a correct
answer to this statement: Trying to time the market,
to get out before it goes down and in before it goes up is (a)
a smart thing to do, or (b) not a smart thing to do?
Most people think that it’s not a smart thing to do–to try to
time the market. Only 11% said yes to that.
But then I asked another question, do you think trying to
pick mutual funds, trying to find a mutual fund
that will beat the market is a smart thing to do or not a smart
thing to do? Most people think it’s a smart
thing to do. What I think the mutual fund
industry has turned into, largely, is a stock picking
industry, not a portfolio diversification industry.
What most people are doing when they go into a mutual fund is
they’re trying to find smart people who will beat the
market–who will pick those stocks that will do well.
The mutual fund theory that we gave last time said,
no the mutual fund is just supposed to be diversifying for
you. In fact, the truth is somewhere
in between. Most mutual funds are providing
some diversification service and they’re also trying to beat the
market. Finally, I just want to say
that I’ve been talking mainly about the U.S.,
but mutual funds have been growing in importance around the
world. There was a recent paper by
Khorana, Servaes, and Tufano, that looked at what
it is that explains which countries have had rapidly
growing mutual fund industries. They found, not surprisingly,
that it tends to–mutual funds have been growing more rapidly
in countries that have stronger securities laws and
institutions, especially laws that protect
individual shareholders rights. Also, mutual funds have been
growing more in countries that have higher level of education
and a higher level of wealth. They also grow more in
countries that have institutional structures that
encourage investing in mutual funds, such as pension plans.
I think it’s a trend around the world that we’re going to see
more and more mutual funds and I think it’s a good thing.
I think they will help us to diversify our risks.
Anyway, I want then to move on to the topic of today’s lecture,
which is insurance. Insurance is the other side of
the risk management institutions that we have.
Insurance evolves separately from securities.
It’s long been a different industry, but the principles are
the same. The principles of
insurance–the fundamental powerhouse–is the principle of
risk pooling. Insurers, just like mutual
funds, are providing risk pooling for you.
Risk pooling means they put a lot of people with independent
or low-correlated risk into a pool and reduce the risk for the
whole pool. They have to contend with
something called moral hazard, which is the risk that people
will be affected by the fact that they’re insured and do
something bad. The classic moral hazard
problem is the problem that you give fire insurance on a house
and someone burns down the house in order to collect the
insurance. They also have to deal with
selection bias. What this means in the
insurance context is that if you offer insurance policies you
will tend to attract people who are higher risk.
If you offer life insurance, you have life tables which give
you probabilities of dying at various ages,
but that’s for the general population.
All the sick people will come to you to buy life insurance and
they will turn out to have a higher death rate than the
population at large. These are the problems of
insurance that we have to deal with.
I just want to review the mathematics of insurance.
This is actually just, in part, just a review of what
we talked about in the second lecture.
In the ideal world, if you have independent risks,
under the independence assumption,
the probability distribution for the number of insurance
contracts that you will have to pay on follows the binomial
distribution. x is the number of
accidents–let’s say this is some accident insurance–the
probability of having–n is the number of policies that
you’re writing. If you’re going to have
p as the probability of an accident, then the binomial
distribution gives you the probability of having x
accidents out of your n policies.
We had that before, that is p^(x)(1-
p)^((n-x)) n!/(x!
(n – x)!) That is the binomial distribution and it
allows you to calculate the probability of any number of
accidents. The mean proportion–The mean
of x/n, is equal to p.
If x is the number of accidents, the mean number of
accidents divided by your number of policies is given just by the
probability, but the standard deviation of
x/n is equal to the square root of p(1
– p)/n. That is the–that gives you the
mean and standard deviation of the proportion.
To actually apply this it helps to go to something called the
normal approximation to the binomial,
because it’s kind of difficult to compute this formula.
There’s an easier formula and you assume that the binomial
distribution is really a normal distribution with a
mean–I’m sorry, the proportion of accidents
x/n follows a normal distribution with mean
p and standard deviation given by this.
That’s the whole theory that I have here;
it’s simple. Maybe I should make a
bigger–let me do it here, I can fit it in here I think.
I’m going to draw an example. I’ve got it plotted out and
it’s on the website, but it’s a very simple example.
I have the case where p=.2, so the probability of an
accident is 20%. This is a significantly high
probability of accidents. Can you see this?
Let’s do this from 0 to .4. If you wrote only one policy,
what’s the probability distribution of
x/n? Well, it has two possible
values. It could be the one person
doesn’t have the accident or does have the accident.
So, if n=1 we have an 80% chance of no accident
and–let’s make this 1 not .4–then a 20% chance of
x/n=1. I’m plotting–this is the
probability of various values of x/n.
If n=1, x/n can take on
only two values: 0 or 1.
It takes on the value 1 with the probability of 20% and the
value of 0 with the probability of 80%.
I didn’t use the normal approximation there–that’s
obvious–I used the binomial itself.
Let’s go to the case where n=100.
You can see this okay? Then, if n=100–now
I’m going to label this x differently, I’m now going to
show the normal bell-shaped curve and I’m going to do this
from 0 to .4. Maybe I do have to make this
bigger–can you see this all right back there?
0 to .4, so .2 is here in the middle;
that’s .2. For n=–now I’m going
to do n=100. What is the mean?
Well, the mean is always .2 no matter how many policies you
write, so it’s going to be–we’re going to have a normal
distribution centered on .2. This .2 is not very readable
here. What is the standard deviation?
It’s the square root of (.2 x .8)/100.
So that’s .16/100; so the standard deviation is
.04. What does the curve look like?
It’s a bell-shaped curve that looks about like that. I didn’t draw that very well,
let’s do it again. I should be able to do nice
bell-shaped curves, but it’s harder than it looks
standing up here; so that’s your bell-shaped
curve. With 100 policies,
they can’t really count that accurately on having 20% of the
policies paid. There’s still substantial
insurance risk because the–it could easily be only 15% or it
could be 25% of the policies that end up paying.
So, the insurance company with one hundred policies would have
still substantial risk. It’s much better–the
uncertainty about x/n is much lower
for the case n=100 than it was for the case n=
1, but it’s still there.
Now, I want to draw–what if we write 10,000 policies?
What is the probability distribution for
x/n in that case? Well, you can see you’re going
to be dividing not by a hundred here, but by a hundred times a
hundred, so it’s going to reduce this from .04 to .004.
So the normal–the bell-shaped curve in that case for the
x/n is going to look something like this.
I’ve got it plotted here; it’s even more steep than that.
That’s a bell-shaped curve, but it’s centered exactly on
.2. Number of policies doesn’t
affect the means but it affects that standard deviation,
so it becomes very collapsed and this is the basic core idea
of insurance. You have to be a big company to
do it and if you have a big company you’ve exhausted the
risk of–never goes away completely.
If we did a–if we go another two decimals–if we did a
million policies, then we would–this would
almost just be a spike here at that point,
so that’s the concept of insurance.
The idea really took root–the idea, the intuitive idea is that
as you write a large number of policies,
the fraction that will result in accidents becomes closer and
closer to the probability of one accident.
That’s an idea that struck people intuitively at various
times in history, but they didn’t know how to do
these calculations. Historians of probability have
noted that–this basic idea is actually in Aristotle,
the ancient Greek philosopher-scientist,
and I’m going to quote Aristotle from his book,
De Cielo. He says–he’s talking very
generally here but I think it’s just–he’s really talking about
this. He says, “to succeed in many
things or many times is ‘difficult.’
For instance, to repeat the same throw of
dice 10,000 times would be impossible;
whereas, to make it once or twice is comparatively easy.”
That’s Aristotle talking. It’s exactly this theory,
but he doesn’t have the word probability, which hadn’t been
invented yet. So, he’s using intuitive–he
says difficult or easy–so he says it’s difficult,
meaning the probability is very low, or easy,
meaning that probability is high.
He had the idea but he didn’t have the math.
I think that may be the first-known statement of the
binomial distribution. Well, it doesn’t get very
precise but it has the intuitive concept.
The idea of using this theory for insurance,
it’s–the earliest-known statement of it is a–was in an
anonymous letter written in 1609 to Count Oldenburg;
but that’s not the person who said it, that was the person who
received the letter. Anyway, the person wrote–he
was talking about fires and was proposing that people should pay
1% of the value of the home every year into a fund and then
the fund would be used to replace the home if it burned
down. Quoting this anonymous person
writing in 1609, “There is no doubt that it
would be fully proved, if a calculation were made of
the number of houses consumed by fire within a certain space in
the course of thirty years, that the loss would not amount,
by a good deal, to the sum that would be
collected in that time.” It’s interesting that this
person uses the word calculation;
this person has the idea–this was–1600 was around the year
when probability theory was invented.
So, someone had the idea of going beyond the intuitive
notion that Aristotle mentioned and moving to something that is
calculable. That’s when the insurance
industry was really born. Insurance relies on this theory
of risk pooling but it has to make it work.
I stressed in the third lecture that like any–insurance,
like any other risk management device, is an invention.
Every risk management device relies on a design and the
design is usually complex and has–it all has to work
together. In order for a design to work
well we have to have every component there.
If one component is missing we may have a failure.
All these components have to be compatible with each other and
it has to work according to a plan, which ultimately is
informed by this theory. Insurance as an invention has
to have what things? It has to have a contract
design; that would be a document,
which is a contract between the insured and the insurance
company. It specifies–what does it
specify? It specifies what risks are
covered, exclusions–some risks are not covered.
Those exclusions are carefully designed in regard to moral
hazard and selection bias. There has to be the
mathematical model, which I just presented,
but it may be more complicated. There has to be a form for the
company. There could be a corporate form.
There’s the insurance company, could be a corporation and it
could be either a non-profit corporation or a for-profit
corporation owned by shareholders.
Or the insurance company could be a mutual insurance company,
in which case the insurance company is owned by the insured.
Then you need, as well–you need government
regulation because the insurance companies don’t seem to exist
without it. There have to be regulators
that are at least verifying that the insurance company is doing
what it says it’s going to do. One thing that the government
has to do is reassure that reserves’ requirements are met.
The company has to have enough money on hand to pay out in the
case of default. One way of classifying
insurance companies is that they can exist as either multiline
insurance companies or monoline insurance companies.
A multiline insurance company insures many different kinds of
things and doesn’t confine itself to one thing.
If a company were just a fire insurance company it would be a
monoline insurance company. They’re essentially more risky
and more–regulators have to watch them more because they’re
standing at higher probability–a monoline–of some
major disaster. In contrast,
a multiline kind of insures itself.
We hear a lot about monoline insurance companies in the
newspapers today. I don’t know if you were
reading about the subprime crisis, but we are living in a
time, at this moment,
of a financial crisis called the subprime crisis.
That naturally hits monoline insurance companies more than
multiline insurance companies because they are more
specialized and more vulnerable. I’ll come back to this
distinction in a minute. I want to talk about certain
kinds of monoline insurance companies that–the biggest–one
category is property and casualty.
These are insurance–or P&C–these are insurance
companies that insure the value of a home or a business or an
automobile. Another kind of monoline
insurance is health insurance company that merely insures
people against health costs, and another important category
is life insurance. Life insurance is–what it
insures is a beneficiary against the death of an insured.
The classic example–the most important example is families.
You would buy life insurance on both the husband and the
wife–you might do different amounts;
it’s done by families with young children to protect them
against the economic cost of the death of one of their parents.
You need insurance on both parents because both parents are
contributing to the success of the children.
These are big industries. The total assets of property
and casualty in 2007 in the United States–and these are
assets–the property and casualty was $1.4 trillion
dollars and life insurance was $4.9 trillion dollars.
Somehow I’ve missed getting the health, but that would be
another big insurance. These are hugely important
institutions in our society. What our–what does–let me go
to property and casualty. What do they insure?
Actually, the most important things that they insure are
automobiles. The total of the premia
collected on automobiles is about–is much bigger,
like five times bigger, than of the premia they collect
on homeowners. Automobile insurance is
collision insurance; homeowners insurance
is–it used to be called fire insurance but now they’ve
extended it to include so many other risks.
It depends on the particular policy–what it includes.
It might include risks, as well, of personal liability
if someone injures themselves on your property or risks of
storms, of hurricanes,
earthquakes, everything else;
so we call it homeowners’ insurance.
Actually, the automobile insurance is more important than
the homeowners’ insurance, even though homes are so much
more valuable, because–I think that’s because
cars move and they drive around and they bump into each other.
Homes just stay where they are, so we have far fewer accidents,
so they don’t have to charge as high a premium for the
insurance. These insurance contracts have
come across gradually through time as we develop the theory
and–I’ll talk about the growth of insurance and about some of
the components of it. The real insurance industry,
as I mentioned before, began in the 1600s with the
invention of probability theory and with the invention of life
tables for–the invention of actuarial science;
but, it grew slowly. I think the reason that it grew
slowly was that insurance is a very sophisticated concept.
In order to explain it–I had to write down the binomial
distribution to explain it properly.
For most people, that’s a difficult concept
and–I think I may have mentioned some of this before,
but let me give this history. Insurance was invented in the
1600s but it did not proliferate fast, it proliferated only very
slowly. Some of the important
figures–there had to be certain inventions to make it work.
Again, I’m repeating a theme that is in my book,
New Financial Order, that financial innovation and
insurance innovation are successions of inventions and
each invention propels the idea more forward.
It’s like we have laws of thermodynamics that underlie the
use of engines, but you can’t just go from the
laws of thermodynamics to an automobile;
there are a million steps along the way.
If you look at the history of engines there are discrete
advances when people were able to apply the theory more and
more. Well, in insurance there’s a
similar list of inventions. Morris Robinson was head of
Mutual Life of New York in the 1840s and he got the idea of
highly-paid life insurance salespeople.
The idea was that it was difficult to sell people on
insurance. Back in the 1840s,
life insurance was very important because the average
expectancy of life was only something like forty-five years,
so that meant parents were dying left and right.
What would be the probability that a married couple would
live, both of them, to the time when their children
were grown? Well, it was fairly low, right?
If the average age of death was something like forty-five,
maybe a fifty-fifty chance of–high risk–of one of the
parents dying. You should think that people
would really want life insurance but they were not buying it.
Why was it? First of all,
they were dying in such numbers;
p was so high that it was expensive.
They have to pay–the premium has to cover the costs,
so it was tough to get someone to buy life insurance even
though they really needed it–it was such a good idea for them.
It was partly because there was a psychological resistance to
it. This is still very much alive
today. I was standing at the World
Economic Forum at one of our lunch things and a young woman
approached from Swiss Re, which is the Swiss Reinsurance
Company, and she said she wanted my ideas on how to sell crop
insurance in Africa. She said, we have it now at
Swiss Re and, of course, The World Bank
sponsors crop insurance for farmers.
There are some very poor areas in Africa where farmers really
run the risk–if their crop fails it could be really bad;
they would be approaching starvation.
So, wouldn’t you think that farmers would want to buy crop
insurance from this Swiss company?
Looks like a good idea to me, but she says,
we’re having a lot of trouble selling it.
She said, if you talk to the farmers in these rural areas
what do you think they say when you offer?
They say, I can’t afford the insurance.
Well, they’re not thinking right.
The whole idea of insurance is that you take from your good
years and you move it into your bad years so that you make it
through all your years. So, you’re having a good year
this year maybe–it looks that way now–and you think you can’t
afford it, but just think how bad it will
be if it’s a bad year. Then you really won’t be able
to afford to stay alive. But she says, they didn’t;
some of them respond but a great majority of them don’t.
I think it’s because of a psychological aversion that
people have to thinking about insurance;
it’s just unpleasant. Life insurance is actually an
insurance against one of you dying;
it’s a very unpleasant topic. If someone comes to your home
and says, I would like to sell you life insurance.
You think, some other day, I don’t want to talk to day
about the probability of one of us dying.
So, it was a tough sell. Morris Robinson,
however, realized that some people are very talented
salespeople and they’re probably talented at other things as
well. It may sound like a small
improvement but he just had very highly-paid insurance salesmen
and they were paid as long as people kept their policy.
That motivated insurance salesmen to form long-term
relationships with the families that he was insuring and to keep
them from canceling their policies.
He got talented, respected members of the
community who people admired to become life insurance salesmen
and he had to pay them enough so that they would stick with the
job; then it finally worked.
It may seem crazy but–it may seem like a modest innovation
but it actually was–it was an important innovation.
I don’t know how you think of life insurance salespeople,
but they have been pillars of the community.
They are people that you–the community–trusts,
that people are willing to let into their home and discuss
things like death; that was an innovation that
came in then. The other innovation was by
Henry Hyde and he was at another insurance company,
Equitable Life, and that was in the late 1800s. What he invented was an
insurance policy that had a cash value and that’s another–it’s a
cash value on the insurance policy.
That is, the insurance policy doesn’t just insure you against
death, it also builds value over the years.
This invention was very important because it stopped
people from canceling. The big problem life insurance
had was people would buy it and they’d pay for several years and
eventually they would think, well we didn’t die,
we’re losing all this money, let’s just cancel the
insurance. It especially happens–the way
life is in real families is: you start getting accustomed to
a certain style of life; you start spending more and
more money; there comes a time when you
have a little crunch and you’re a little short on money;
you’re casting your net out for some way to come up with some
money; and canceling your life
insurance was a good idea. It turns out that if you make
them forfeit their cash value on canceling, then they won’t
cancel. Both of these were ideas that
were copied all over the world–that’s the way inventions
are–so a lot of insurance policies today have cash values.
The third thing I was going to report was that sociologist
Viviana Zelizer wrote a book about life insurance sales in
the nineteenth century. She found that there was a lot
of resistance to the purchase of life insurance in the nineteenth
century. She tried to study it and tried
to figure out how it was that life insurance became more and
more important over that century.
One of her conclusions was that life insurance seemed to be
opposed quite a bit by women, nineteenth century women. Why didn’t they like it?
You would think that any rational woman in the nineteenth
century would reflect on the fact that there’s a significant
probability that her husband will die of something while you
still have children. Why wouldn’t they want it?
Well, what she found was that life insurance salespeople were
going to families and trying to sell them on life insurance by
explaining the concept and they would say something like I did.
Maybe they didn’t write down the binomial distribution,
but they explained the idea of insurance and it didn’t sound
right to the typical nineteenth century American woman.
I suppose it wasn’t just America, it was a worldwide
problem. The reaction that salespeople
would get from women was, you’re giving me some
probabilities or something, you’re asking me to–it sounds
like you’re asking me to play in some gambling thing where I win
if my husband dies; it doesn’t sound right to me.
I think, in fact, a lot of women would say
something like, I put my faith in God and I
think I might bring down God’s wrath if I did such a thing as
to bet that my husband is going to die.
So, she would refuse to take part in it;
it doesn’t sound right to her–I’ll trust in prayer and
other things. It didn’t work and you couldn’t
sell them on it. What Zelizer reported was that
some life insurance companies surmounted this problem by
changing the pitch, by telling their salespeople,
don’t try to explain probability theory to these
people. What you have to do is come
across differently. The thing that she said they
started doing was to pretend, in a way, that they were
missionaries with a gospel and the gospel was insurance.
They would tell these women, you know, if you buy life
insurance on your husband, then should anything happen
your husband can love and protect you from beyond the
grave. That sounds good–it worked and
these little things–they may seem like little things but they
are technological advances. People in a profession over the
years learn more and more about how to manage the public’s
expectations and get them to actually purchase insurance.
I want to talk about government regulation of insurance because
I said that was an important aspect of insurance.
The regulators do many things, but insurance regulators,
most of all, are concerned with capital
adequacy. Insurance now,
in the United States, is regulated not by the federal
government but by insurance regulators in each of the fifty
states. It’s very hard to summarize
insurance regulation in the United States.
Why is it regulated separately by the fifty states?
That’s because it started out that way;
when the federal government wasn’t involved in those sorts
of things and somehow we never had the–in fact,
it’s talked about–we should–maybe we’ll see it in
the next ten or twenty years. It’s talked about that there
should be a federal government regulation of insurance but it’s
actually state regulators. It makes it–this is a handicap
to the U.S. insurance industry.
Other countries have it centralized but in the U.S.
it’s split up over fifty states. It’s very hard to start an
insurance company because you’ve got to meet the requirements of
all fifty regulators. The most important thing that
these do is: they specify what reserves insurance companies
have to hold. So they–in other words,
the insurance company–doesn’t trust the insurance companies to
do the calculations like I showed with the binomial
theorem. They want to make sure that
there’s a significantly high probability–sufficiently high
probability–that even if they get a bad draw and there are a
lot policies that require paying out,
that these reserves will satisfy.
An insurance company must hold the reserves;
it can hold more and that’s called the statutory surplus.
The reserves are an accounting entry–it’s how much they are
required to hold, but the companies will hold
more than that, typically, in order to protect
themselves–more than is required–and their policy
holders–more than is required by the regulators.
Occasionally, you get into problems with
reserves. Before I get to that let me
just say–I want to just mention a few types of insurance that
are important. Let’s talk about life
insurance–and we’re talking about kinds of policies that are
around today. The simplest form of life
insurance is called “term insurance.”
This is insurance that you pay each year for insurance in that
year and it does not build a cash value.
I can buy insurance for myself this year and I can stop and if
I–there’s nothing gained–or next year I would just have to
pay it again. “Whole life” is more
complicated because it builds a cash value according to a
schedule and there is both non-participating and
participating. With non-participating–with
participating, you are participating in the
portfolio outcome that the insurance company is
experiencing, so you have some uncertainty
about your cash value. There’s something called
“variable life,” which refers to a life insurance policy where
the policyholder can make decisions about the investment
of the money in the whole life policy.
It’s not just taking what’s given by the insurance company.
There’s something called “universal life”–these are all
explained in Fabozzi, et al.
It’s a whole life policy that gives the policyholder
flexibility over the insurance premiums.
You can pay more into the cash value in one year and less in
another year, as long as you keep paying a
minimum amount. There are also survivorship
policies that will pay, for example,
the second to die–there would be a policy if a husband and
wife get it, then it pays out when the
second of them dies. Also, about regulation,
I want to mention the NAIC–it’s a very important
institution–that stands for National Association of
Insurance Commissioners. This is something that’s very
important in the United States because, as I mentioned,
insurance regulation is divided up over the fifty states and the
problem that that creates is–it’s a nightmare for
insurance companies because every different state has
different laws. Recognizing this problem,
the insurance commissioners in the various states have decided
to form an organization, that’s the NAIC,
and they hold regular conferences.
At these conferences, they decide on a recommended or
uniform insurance regulation policy and it acts a little bit
like a Congress. They make laws but the laws are
not binding on anyone, they’re just suggested laws or
regulations. It’s an effort to try to get
the complexity of fifty different regulators uniform.
As I say, when the NAIC decides on some regulation,
it’s only a recommendation to the separate state regulatory
commissions. But of course,
it has a lot of force because all fifty states have met–the
commissioners of fifty states have met and hashed it out.
So, most states would essentially adopt what the NAIC
says, otherwise–they’re not going to figure it all–they
can’t claim to beat–it would be rational to do that.
You wouldn’t think that we would rethink all of these
things ourselves and then have it different in our state.
The NAIC is kind of a quasi-regulatory body–or a
quasi-government–it’s not government because it has no
authority and yet it acts almost like a parliament where these
people get together and decide on things and they end up as a
force of law. Another important–-there
are some milestones in insurance that I want to mention.
Then I’m going to come back to finally concluding with problems
that we see. Of course, the problems are not
damning; they’re problems that reflect
the progress yet needed to be made in insurance.
I just want to mention the Gramm-Leach-Bliley Financial
Modernization Act of 1999. What this did is it allowed
banks to offer insurance or to ally and join in and merge with
insurance companies. Before that,
as I said before, insurance is really the same
thing as securities. They’re both based on risk
management and pooling of risks. Diversification and pooling are
really the same thing but we had a separate set of institutions.
People never thought of insurance as the same as a bank,
but since this is relatively recently–that was only nine
years ago–at this time, it means that we are now seeing
an expansion of our banking system in the U.S.
to become insurance-related. It’s different in other
countries; in Europe, they’ve had
universal banking, which allowed banks to offer
insurance for a much longer time but in the U.S.
it’s kind of a recent innovation.
I want to conclude with problems, which may sound kind
of negative but this is–I don’t shrink from negativism.
It’s not really negativism; it’s talking about what should
be done. I want to come back to–I told
you that monoline was in the news a lot lately.
So, what are we talking about? I don’t know how much you read
about the–right now we are going through a major financial
crisis, which started in the U.S.
but has spread over the world and this is called the subprime
crisis. Subprime refers to mortgages,
which is not the subject of today’s lecture,
but a subprime mortgage is a mortgage issued to a borrower
who is not considered prime–not a good risk.
They are borrowers that are thought, by the various models,
to be likely to fail to pay on their mortgage and to have to be
foreclosed; often low income,
but also they’re people with poor credit histories.
The crisis that we’re in now–and this is very important,
you’re living through it and we’ll see how it pans out.
The subprime crisis is happening today because home
prices are falling and with falling home prices more
subprime borrowers are failing to pay their mortgage.
Default rates are shooting up and valuation of securitized
subprime mortgages have crashed and it’s throwing turmoil all
over the financial community. We’re seeing what they call
systemic effects–that’s a very general term that goes beyond
insurance. By systemic effects,
I mean something that affects the whole financial system.
When you do these calculations, which I just did for
insurance–I was talking about insurance companies–assuming
that risks are independent, everything’s independent and
we’ve got it all figured out. Underlying it there were other
things besides just the calculation of their accident
rates that had certain assumptions built in.
The failure of the assumptions in many different industries can
create systemic effects. So what happened?
I’m going to talk about a particular line of monoline
insurers that you may never have even heard of because they don’t
deal with the general public. These are municipal bond
insurers. These are private
companies–insurance companies–they’re monoline
because they look only at a certain class of risks and not
all risks that they insure. They deal with,
principally, state and local governments who
are issuing bonds to raise money for their activities,
like New Haven, for example,
or any other state or local government.
We refer to the bonds that they issue as municipal bonds.
Now, the problem is that if you invest in bonds from some state
or local government, they might not pay you back;
they can just go bankrupt. Cities go bankrupt and they
just can’t pay, so you as a buyer of these
bonds feel reluctant to invest in them.
In order to make their bonds saleable to the general public,
city and state governments go–or any region or local
governments go to the municipal bond insurer companies and they
have names–the big ones are MBIA,
AMBAC, FGIC–that’s not FDIC, it’s not the Federal Deposit
Insurance Corporation, it’s FGIC–and there are others
as well. What these do is they insure
the bond–they insure the investor against the
municipality failing to pay on the bond.
Lot’s of investors won’t buy a municipal bond unless it’s
insured. I’m talking about systemic
effects. So, these insurance companies
have reserves on hand and they invest these reserves in
something. What do they invest them in?
Well, one of the things they’ve been investing them in are
subprime mortgages. Now, you can see why I’m
talking about systemic effects. The housing sector in the
United States, you might say,
is completely different from municipal bond insurers;
but when the housing sector starts going down,
people start defaulting on their mortgages.
Then, the value of the subprime loans that the municipal bond
insurers own starts to go down. Now, these guys have never
failed to live up to their guarantees–they’re doing just
fine–but people are noticing that their portfolios are going
down, so their surplus–statutory
surplus–is going down, so we’re starting to worry
about these companies. Notably, it was January 18,
a rating agency, Fitch, lowered AMBAC from
AAA-rated to AA-rated–that was a big news story.
You might not have caught it, right?
It was a big news story for the people who work in the municipal
government. You say, oh oh,
this is bad news, if this is the only–if only
one rating agency does that, okay, but–and it’s only AMBAC
that’s been down-rated, but you starting wondering.
So, reporters start calling the other rating agencies–Standard
& Poor’s and Moody’s and others–and say,
are you going to down-rate these guys?
You read in the newspaper their interpretation of what’s going
to be said and so far the other rating agencies haven’t;
they’re not saying anything. Now people are suspecting that
these are going to get down-rated.
Well, if they’re down-rated, then nobody trusts them anymore
as insurance companies. That means that the municipal
governments will find it hard to issue bonds to continue their
activities, so the municipal governments
might have to shut down some activities, like fixing the
roads, or bridges, or building
schools–whatever the cities do with their municipal bond money
that they raise from the municipal bonds.
You can see how things are feeding through the system:
it starts out in one thing, it goes to the insurance
companies, and then it goes to the municipal governments.
All that kind of thing might put this economy into a
recession. It’s not just municipal
governments, but a lot of different aspects of our
financial system are going to be touched by a crisis that’s
spreading from one segment to another.
That’s where this stands right now.
The latest thing–well, now this is a crisis–a
municipal bond crisis which is unfolding.
The New York state regulators have been trying to get
companies to subscribe more capital to the monoline
insurance companies to bolster their capital so that they have
more money to pay out; that will prevent any more
lowering of their ratings. The other thing that’s happened
now is that Warren Buffett said he wants to get into the
municipal bond insuring business.
Well, he’s just a business person coming in.
So, there will be new municipal bond insurers appearing that
will take up slack and we’ll be all right but the problem is
that we may have a crisis for a while.
The other thing that I want to talk about–and I guess I’m
running out of time–was about management of disaster risks.
We have another insurance crisis developing now because
of–well, a very important source of
insurance risk that’s developing has to do with the rising rate
of hurricane damage that we’re observing in the east coast of
the United States. Notably, we saw Hurricane
Katrina that caused huge property damage in–a couple of
years ago–that, kind of, tested insurance
companies again. The risk is that global warming
will make hurricanes more common and I guess I’ll have to–I’ll
just conclude that. I’ll talk a little bit about
that–I’ve gotten through almost all of this–next time.
The only other thing–then next period we’re going to talk–on
Friday we’re meeting again and we’re meeting again about–this
time it’s about efficient markets and I want to talk about
the evidence for efficient markets and against it and that
will lead you into your third problem set about efficient
markets. That’s not coming for a little